Symmetrical to Asymmetrical Model Transformation for I/Q Imbalance in Zero-IF Transceivers

Abstract

Two main types of modeling are considered in the $I/Q$ imbalance problem literature: symmetric models and asymmetrical models. Specifically, the symmetric models are applied in the $I/Q$ imbalance testing area, while the asymmetrical models have applications in the $I/Q$ imbalance calibration area; each model offers advantages in their respective areas, with applications in real systems. In this paper, a study of both types of modeling was carried out, covering main characteristics, advantages and differences between these models. Symmetrical and asymmetrical models in matrix form are developed for transmitter and receiver bearing on a zero-IF architecture, taking into account the main sources of imbalance in the analog front end (AFE). Based on these models, a matrix model transformation methodology is developed, which is capable of transferring the $I/Q$ imbalance characteristics of a symmetric model into an asymmetric model. Simulation results confirm the proposed model transformation through a performance comparison of four $I/Q$ imbalance testing algorithms in the sense of the AFE imbalances estimation, allowing a fair comparison under the same conditions of imbalance.